Document resource
Introduction Splenic infarcts and abscesses are a recognised sequelae of infective endocarditis (IE). Splenic abscesses were incorporated into the updated Duke guidelines; however, this has not been externally validated, and splenic infracts are not currently included the diagnostic algorithms for IE. We therefore determined the sensitivity and specificity of splenic lesions including splenic infarcts and abscesses in patients with suspected IE and investigated the impact of their inclusion.Methods This is a retrospective diagnostic accuracy study using data from 496 patients with suspected IE who were referred to the Leeds Teaching Hospitals Trust IE service between 2014-2017. The modified Duke criteria from 2000 was used as the gold standard reference test to categorise patients into Duke ‘definite’, ‘possible’ or ‘rejected’ groups for comparison. To calculate sensitivity and specificity Duke definite cases (n= 169) served as true positives and Duke rejected (n=253) served as true negatives to against which splenic infarcts and abscesses were compared. Duke possible (n=74) and Duke rejected (n=253) cases were then used to assess the impact of inclusion of splenic abscesses (n=10) alone and then when combined with splenic infarcts (n=49) on the Duke categorisation.Results Using the Duke clinical criteria as the gold standard comparator, splenic infarcts and abscesses collectively had a sensitivity of 24% (95% CI 18 to 31) and a specificity of 98% (95% CI 95 to 99) in those with clinically suspected IE. Splenic abscesses alone had a sensitivity of 5% (95% CI 2 to 9) and a specificity of 99% (95% CI 98 to 99) in those with clinically suspected IE. Inclusion of splenic infarcts and abscesses as a minor criterion in the Duke criteria would result in a reclassification of 3% of those rejected due to low Duke score to possible and 4% from Duke possible to Duke definite.Conclusion Splenic infarcts and abscesses are an insensitive yet highly specific indicator of IE in patients referred with clinically suspected IE. The addition of splenic infarcts to diagnostic criteria may further increase sensitivity without compromising the specificity, making both useful diagnostic markers for IE with clinical value.